In the world of tomorrow not only you can watch movies, but they are for you

When you are in a dark movie theater, your reaction to what is happening on the screen often go unnoticed by others. Here you wide open the eyes in case of an unexpected plot twist, literally Bouncing in my chair from scary scenes or let the tears under melodramatic music – all of these emotions are likely to be aimed no more than the plastic back of the forward seat. But only in the case that this theater is not equipped with a "computer vision". If you ever get a chance to visit one of these, then rest assured as you watch the film, will look out for you.

A Computer program, whose main objective lies in the fact of what is written above, developed by Silver Logic Labs. Its CEO Garima Ham is an expert in applied mathematics, specializing in number theory. He worked for many years within the walls of giants such as Amazon, Microsoft, and Harvard medical school, engaged in various issues related to the human consumer nature. However, his main interest has always been to understand how artificial intelligence (AI) can help to better predict one of our most difficult features of human behavior.

Exploring AI algorithms to analyze human responses to the perception of different types of media, Jamon realized that it could benefit not only scientific, but also commercial. His system works in the following way: AI-algorithm monitors the audience that watches the film, noting the emotion in the faces of the people, manifested even through the most subtle changes (so-called micro-expressions), and then generates on the basis of this information the necessary data for further analysis.

Jamon seriously delved into similar studies about three years ago and, although the results of his work for the media industry look still very damp, great interest in this topic, they are definitely attracted. Largely because they promise reliable results, as the behavior of people can be as predictable as the operation of the software, at least in the eyes of the Ham. AI in turn, and used computer vision techniques will help to collect information about how people react to certain movies and TV shows, much more effective than any fucus. As soon as the AI train to collect the desired dataset, it can provide a quick, consistent and detailed analysis of this information. For those areas of the industry, who have to work with these data – the news is just wonderful. Indeed, thanks to the more competent and effective analysis of human desires and preferences, they can improve their products and services, bringing the money.

Ratings – one of the most important and significant aspects in ensuring the success of a particular television program or movie. To the surprise of those who watched the process, the current version of the software Ham were able to predict the Nielsen ratings, Rotten Tomatoes and IMDB with accuracy from 84 percent to 99 percent. The difference in figures stems from the fact that some of the objects of the ranking are "multimodal," that is designed for a wider audience, so they usually are harder to predict. In General, when it comes to TV, is in itself an attempt to predict the popularity of a particular transmission can not fail to impress.

"When I started, everyone told me that I will never be able to predict like that, because no one can" — shared Ham in conversation with the portal Futurism.

But with mathematics, nothing is impossible. After all, according to Ham, with the help of mathematical methods we can mention the many nuances that mark without using them is simply impossible.

"We took the emotional reaction to visual and audio stimulation of the audience and convert them into digital values. And when something takes the form of a digital value, then sooner or later it takes the form of the desired equations, a task which in our case boils down to know how do you liked (or like) something or other show" — says Hamon.

Researcher reports that there is a wide statistical analysis, but refuses to divulge any details about what the equation he uses to calculate, pointing to the fact that in this way he tries to protect the "secret ingredient" of his program.

outside of the entertainment industry

High efficiency of AI in the prediction of human preferences, prompted Ham to explore other areas in which his program could be effective. For example, in determining whether a person speaks the truth or not. Like the polygraph, the AI could map data indicating stress with a standard set of values and on this basis to determine whether a person is lying or not. To test this idea, the Ham used AI algorithm to perform the task of identifying emotions in people, based on the relatively low video quality. For the study he used rollers news Agency CSPAN, as well as shooting footage from the press conference of the President of Donald trump.

At a time when the truth can be questioned, the separation of truth from falsehood can be critical. However, the system can also be used in situations when it is literally life and death. For example, in medicine, where it may be necessary to accurately determine the level of pain experienced by the man, so doctors can find more effective means of treatment.

Ham notes that the system can be useful, for example, in situations where it is necessary to determine the stroke. Despite the fact that the medical staff usually undergo the necessary training that allows them to identify the signs of a stroke, often there are times when missed the so called mini strokes (or transient ischemic attack if scientifically), which is very often followed by a massive stroke, affecting a large area of the brain. Computer vision AI could determine microprismatic stroke or even symptoms or hints of impending diseases before they actually manifest in the patient. In this case, medical personnel would be able to react to the situation, and may even take steps to avoid further development of a more serious stroke.

But will it actually work? Hamon believes that this is possible. The researcher is convinced that the AI can become so sensitive that it will be able to define such barely noticeable and rapid changes in the body. When the system was tested at different audiences, the researchers had to take into account the fact that some people receive prescribed medicines, some of which, for example, have side effects such as increased pressure or a barely noticeable muscle spasms. Perhaps you will not be able to notice these changes in another person, however, AI easily identify, but at the same time, it may make the effects from taking the drug for the manifestation of stress symptoms. It also must be considered.

Powerful tool

Many believe that intelligent machines will be deprived of bias, however, we should not forget that these machines will create people, and these people, developing and interacting with AI, can unknowingly give them their own prejudices. With the development of AI the consequences of the accumulation of these biases will become more pronounced and eventually can affect various processes of information processing, for example, when identifying a particular individual, or the collection of social data of a person contained in the networks of public services.

As technology of facial recognition is already beginning to penetrate our personal lives, many are beginning to think and pay more attention to ethical issues, expressing concern about the potential bias of these algorithms. The same idea that technology can exist without any degree invested in their biases, is very controversial. Serious doubts about her puts only a single argument: the efficiency of AI will depends on the data that was invested in it, and among these records could contain information that was originally painted a biased attitude of the man who created this system. With the progress of development of AI, creating machines that are actually able to learn, we need to develop restrictive measures to protect us from situations when these machines will be able to learn from us much more than we intended to teach them.

However, the same Ham is confident that his algorithm is completely impartial, at least as much as possible. Its computer system only interpreterpath signs of human behavior, no matter what type of face or body is in its field of view.

"I'm native American, and we should recognize that there sometimes occur things that can aggravate the surrounding environment. For example, you can start to get nervous when your back goes police. However, I believe that such technology in the future will be able to completely eliminate this factor of nervousness. If you are doing anything wrong, the computer will alert police that you are not doing anything wrong. Personally I in this case, it will sense the increased level of your own safety and protection from police brutality, knowing that the computer will be able to carry out this level of threat assessment".

In any case, the Hamon does not care about the interpretation of the results of the data analysis algorithm that he created. At the same time, despite the confidence in his algorithm, he also recognizes and ...